trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.951219512195122,
  "eval_steps": 8,
  "global_step": 20,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.78,
      "grad_norm": 17.967552185058594,
      "learning_rate": 0.0003333333333333333,
      "loss": 1.3774,
      "step": 8
    },
    {
      "epoch": 0.98,
      "eval_accuracy": 0.525,
      "eval_f1_score": 0.5204761904761905,
      "eval_gmean": 0.5200265681471932,
      "eval_loss": 1.061352252960205,
      "eval_precision": 0.521875,
      "eval_recall": 0.525,
      "eval_runtime": 163.6249,
      "eval_samples_per_second": 0.489,
      "eval_steps_per_second": 0.061,
      "step": 10
    },
    {
      "epoch": 1.56,
      "grad_norm": 8.729334831237793,
      "learning_rate": 0.0001111111111111111,
      "loss": 0.6975,
      "step": 16
    },
    {
      "epoch": 1.95,
      "eval_accuracy": 0.6375,
      "eval_f1_score": 0.6376701611641371,
      "eval_gmean": 0.637217982733379,
      "eval_loss": 0.7615417838096619,
      "eval_precision": 0.6380393996247655,
      "eval_recall": 0.6375,
      "eval_runtime": 164.1513,
      "eval_samples_per_second": 0.487,
      "eval_steps_per_second": 0.061,
      "step": 20
    },
    {
      "epoch": 1.95,
      "step": 20,
      "total_flos": 2.536276286767104e+16,
      "train_loss": 0.9586615085601806,
      "train_runtime": 8394.2659,
      "train_samples_per_second": 0.153,
      "train_steps_per_second": 0.002
    }
  ],
  "logging_steps": 8,
  "max_steps": 20,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 2.536276286767104e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}